Hutchinson — On the Estimation of Oxygen in Water. 143 



large quantity of carbonic acid present ; for, as Schutzenberger points 

 •out, when an acid is present in appreciable quantity, even such, a weak 

 acid as carbonic acid, the results given are invariably too low ; hence 

 this metbod would not be applicable, with any degree of accuracy, to 

 •waters in wbicb a large quantity of carbonic acid is present. 



I made numerous determinations, from which, the two foregoing 

 examples are selected ; but none of them showed any trustworthy 

 results ; in some cases the volume of oxygen obtained being in excess, 

 and in others less than that obtained by the Gasometric method. 



The variability of the results led me to inquire into the source of 

 these discrepancies, and how they might be avoided. 



Noticing that a change of colour in the yellow neutral tint seemed 

 to occur to a greater extent than it should do on the addition of re- 

 cently boiled distilled water, I made the following experiments. 



Distilled water was boiled in a flask fitted with a cork and exit 

 valve so as to avoid contact with air, for over four hours. The appara- 

 tus was prepared as usual with indigo and hyposulphite, the tem- 

 perature being kept at 50° C. The boiled water, which was kept in 

 well-stoppered bottles, was then added in successive portions of 75 cub. 

 •cents, at a temperature of 55° C. On each addition a blue colouration 

 was produced in the yellow neutral liquid, just as if oxygen had been 

 absorbed by the reduced indigo. The amount of colour change was 

 determined as usual by the addition of hyposulphite. The mean of 

 five experiments showed that 2-9 cub. cents, were necessary to bring 

 back the yellow tint. This quantity was found to be equivalent to 

 3"86 cub. cents, of indigo, or 0"201 cub. cents, of oxygen. 



Unwilling to think that this was due to oxygen which had been 

 left unexpelled by ebullition, I boiled recently-distilled water in long- 

 necked flasks, fitted with corks and exit valves, for over five hours. The 

 apparatus was prepared as usual, but in this case water at 100° C. was 

 used, and the body of the apparatus immersed in water kept at 100° C. 

 Portions of 75 cub. cents, of water were removed from the flasks 

 whilst in a state of ebullition, and introduced into the apparatus ; each 

 addition caused a change of colour fi'om yellow to blue. The mean of 

 five experiments showed that 1'25 cub. cents, of hyposulphite was 

 necessary to destroy the colour. 



1-25 cub. cents, hyposulphite = 1*66 cub. cents, indigo = 0-08 

 cub. cents, of oxygen. 



After boiling for over six hours, the water was allowed to cool, out 

 of contact with the air, and in withdrawing portions from it coal gas 

 was aspirated in, instead of air, so as to avoid as far as possible contact 

 with oxygen. 



The same experiments were tried at the ordinary temperature 

 (16*75° C), and the same volume of water (75 cub. cents.) used. The 

 mean of four experiments gave the colour change equivalent to 1"3 

 cub. cents, of hyposulphite = r56 cub. cents, of indigo, or 0-087 cub. 

 cents, of oxygen. 



The same experiments were now made at the temperature O'C, by 



